zlacker

The Lonely Work of Moderating Hacker News (2019)

submitted by capabl+(OP) on 2023-07-28 19:22:10 | 225 points 214 comments
[view article] [source] [go to bottom]

NOTE: showing posts with links only show all posts
◧◩
6. carabi+el[view] [source] [discussion] 2023-07-28 20:59:29
>>belfal+Nh
Yep this happened with my last 3 flagged submissions. All on social issues. Really sad because especially first one listed below I thought would elicit good discussions, somewhat tied other issues like affirmative action.

>>35867458

>>36065735

>>36627969

10. ChrisA+Gm[view] [source] 2023-07-28 21:07:22
>>capabl+(OP)
Annual submission? Forever props to Dang and the mods of course.

Some previous discussions:

A year ago >>30374040

3 years ago >>25048415

4 years ago >>20643052

◧◩◪◨
22. ryandr+jr[view] [source] [discussion] 2023-07-28 21:29:16
>>vector+lp
I find the topics that are most likely to generate "heated" discussion tend to be emotional topics, but also they're more broadly important to society and just more interesting to discuss. That's why I tend to browse in /active[1]. Some JSON parsing command line toolkit re-written in Rust [4 comments]? Yawn..

1: https://news.ycombinator.com/active

◧◩◪
24. dang+Bt[view] [source] [discussion] 2023-07-28 21:43:32
>>carabi+el
I'd say >>35867458 was an interesting submission that probably shouldn't have been flagged, although it's doubtful whether a thoughtful, curious discussion is possible. Usually we end up with people charging in and wielding their priors as a stick, and from that point of view I can understand the flags.

The other two were too sensational and, in the case of >>36627969 , already a heavily discussed theme, so I'd say they were flagged correctly.

◧◩
29. dang+8v[view] [source] [discussion] 2023-07-28 21:52:14
>>motoha+uo
https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...

https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...

◧◩
30. dang+ov[view] [source] [discussion] 2023-07-28 21:53:52
>>gerdes+no
She did do a ton of research, worked through all the materials we sent her (I sent a lot, including a thwack of Kids in the Hall videos, several years-long email threads with specific users, with their permission of course, and god knows what else), and even changed her mind—a thing rare enough to always deserve respect. We took a risk in trusting her; I felt like it worked out, and I'm glad we chose to be open. The article came out fairer than we probably had any grounds to expect. There were bits I could complain about, but that's inevitable. I felt seen and I think Scott did too.

Re Canute, she missed that pvg was being playful, in the context of a longstanding positive connection. I felt bad when I saw that. But we did get an amusing, and properly italicized, About box out of it: https://news.ycombinator.com/user?id=pvg.

Come back, pvg!

◧◩
32. kayceb+3w[view] [source] [discussion] 2023-07-28 21:57:10
>>noir_l+Jn
"When you do things right, people won't be sure you've done anything at all." https://youtu.be/edCqF_NtpOQ
◧◩◪◨
37. dang+Xz[view] [source] [discussion] 2023-07-28 22:21:25
>>1000be+4z
Hey, that's not fair! HN is for everyone's pet topics and non-technical stories are welcome. The only criterion is that they be interesting and gratify curiosity.

I mean, there's a high-ranking thread about a huge oak table on the front page right now: >>36912861 . As there should be.

(You probably already know this but for anyone who doesn't: HN is explicitly not just a technical forum - see https://news.ycombinator.com/newsguidelines.html.)

◧◩◪
40. dang+yA[view] [source] [discussion] 2023-07-28 22:25:44
>>versio+Zy
Political threads often do go that way, and I understand the frustration. We don't want regurgitation—that follows from what we're trying to optimize for: https://hn.algolia.com/?dateRange=all&page=0&prefix=true&sor....

But the question of how to handle politics on HN is not simple. By the same principle of trying to optimize for curiosity, some content with political overlap is interesting and belongs here. The questions are which forms of it, how much, which particular links, etc.. I feel like after 10 years we arrived at a pretty coherent and stable general answer to that. Not that we get every specific call right—we don't. But the general principle has held up.

For anyone wondering what I'm talking about, here are some past explanations:

>>22902490 (April 2020)

>>21607844 (Nov 2019)

and some related points:

>>23959679 (July 2020)

>>17014869 (May 2018)

and there are lots more at https://hn.algolia.com/?dateRange=all&page=0&prefix=false&so... covering this.

◧◩◪◨⬒
44. dang+iG[view] [source] [discussion] 2023-07-28 23:01:38
>>educti+rA
> You also have the temerity to go outside HN and tell people to rewrite posts on other sites - a corrosive form of censorship

I don't understand what you're referring to. Can you explain further?

Edit: oh, I see - you're talking about >>35719273 . Yes, sometimes when startups ask me for help, I make suggestions about how they can change their articles to better appeal to HN readers, or to avoid pitfalls. For example, mentioning one's startup only at the end of an otherwise interesting article (which I guess people do because marketers told them it was a "call to action", or something?) makes many HN readers feel like the whole article was a bait-and-switch, and then they rush into comments to complain that the article is "just an ad" or whatnot. That's what happened in that thread - e.g. >>35718172 and >>35718321 . In such cases I advise authors to mention their startup right at the beginning - an easy fix. I don't think most people would call that censorship! - certainly the authors who take the advice do so freely and say they're grateful for it.

I can't help but wonder if there's something else to your complaint because it's hard for me to understand why that would be objectionable. If you want to explain more I'd be interested...

Btw re "the richest people in SV" - the startup in that case wasn't SV related as far as I know, and certainly not YC related; I believe it's a spinoff from Andy Pavlo's research group at CMU. I wasn't helping them for any reason other than to make the HN thread more interesting and because they emailed to ask.

◧◩◪◨⬒
49. dang+oK[view] [source] [discussion] 2023-07-28 23:29:04
>>nitwit+Vl
That's true. And there's the additional factor that the more familiar a topic is, the more likely the thread is to fill up with generic comments about it. That's what the HN guidelines call generic tangents: https://news.ycombinator.com/newsguidelines.html. Those are probably the most repetitive kinds of thread, and therefore the least valuable.

This phenomenon shows up in most political threads but also in threads that aren't particularly political. The more generic topics have so much mass that they act like black holes that suck in all the interesting discussion (https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...).

◧◩◪◨⬒⬓
55. dang+HO[view] [source] [discussion] 2023-07-29 00:00:21
>>Mounta+jJ
I'm not sure what you mean by failed spectacularly, but if I'm reading you right, then your comment is a good example of a reliable phenomenon: nearly everyone with strong passions on a political topic feels like HN is biased against, and even is suppressing, their position. (In one memorable case, the topic we were accused of suppressing was actually the single-most-discussed topic on HN by a long shot: >>23624962 . That's how intensely these passions work: even the biggest story gets perceived as censored.)

Let's look at the specific topic you mentioned. HN had plenty of discussions about the lab leak theory, starting in late 2020 and all through 2021. I've listed some below; there were others (and of course many more in 2022 and 2023). Some fell off the front page rather quickly but the biggest ones spent 15, 16, 18 hours on the front page.

Everyone's memory about the pandemic has been retroactively revised by now, but as I recall it, the rehabilitation of the lab leak theory in (semi-)mainstream discourse began when Nicholas Wade published his article in the Bulletin. HN discussed that one thoroughly (>>27071432 ) and there had been several major frontpage threads even before that.

An appeal for an objective, open, transparent debate re: the origin of Covid-19 - >>28582290 - Sept 2021 (307 comments)

Scientists who signed Lancet letter about origins of Covid-19, have 2nd thoughts - >>27631560 - June 2021 (36 comments)

The lab-leak theory: inside the fight to uncover Covid-19’s origins - >>27388587 - June 2021 (1062 comments)

Wuhan lab staff sought hospital care before Covid-19 outbreak disclosed - >>27259953 - May 2021 (346 comments)

The origin of Covid: Did people or nature open Pandora’s box? - >>27071432 - May 2021 (537 comments)

Scientists who say the lab-leak hypothesis for SARS-CoV-2 shouldn't be ruled out - >>26750452 - April 2021 (618 comments)

The WHO-China search for the origins of the coronavirus - >>26609494 - March 2021 (209 comments)

Why the Wuhan lab leak theory shouldn't be dismissed - >>26540458 - March 2021 (985 comments)

US raises ‘deep concerns’ over WHO report on Covid’s Wuhan origins - >>26125145 - Feb 2021 (632 comments)

Ensuring a transparent, thorough investigation of Covid-19’s origin - >>25799858 - Jan 2021 (74 comments)

Israeli startup claims Covid-19 likely originated in a lab, willing to bet on it - >>25585833 - Dec 2020 (351 comments)

(There of course were many threads arguing the opposite as well - I'm just listing these because they're the relevant ones for answering the GP. If this post makes you feel like HN was too supportive of and/or too suppressive of the opposite side, please re-read the first paragraph - it seems to work the same way in all cases.)

71. sacnor+331[view] [source] 2023-07-29 01:50:34
>>capabl+(OP)
I dare you: https://duckduckgo.com/?q=Gackle

Spoiler alert: It assume you're looking for a loud, pushy corvid that hops up on your table and insists it has a right to one of your tacos.

84. mitchb+T81[view] [source] 2023-07-29 02:48:20
>>capabl+(OP)
https://archive.ph/V9z3G
◧◩
95. dang+Cn1[view] [source] [discussion] 2023-07-29 05:28:25
>>Frustr+bF
> It seems like HN has become more political in just last couple months.

Users have been saying that almost since the beginning, but I think it's mostly just swings and fluctuations. Lots of past examples here: >>17014869 .

◧◩◪
115. johnny+9B1[view] [source] [discussion] 2023-07-29 08:16:08
>>leonon+Vj
I do. You could even see how quickly discussion evolves when talking about the education system: >>36839983

far from the worst example, but a pretty recent one to show a bit of how bad it can get. I guess I was still (naively) surprised that affirmative action is still such a hot button topic.

◧◩◪
120. johnny+ID1[view] [source] [discussion] 2023-07-29 08:44:49
>>ggm+TQ
>Something happened (in my opinion) between about 84 and 94 which systematically eroded and undermined women's experience in ICT.

you're pretty much on the money: https://www.npr.org/sections/money/2014/10/21/357629765/when...

The reasons speculated are much more boring, though. The personal computer was treated no differently than a Nintendo console, AKA a toy. So they (big coporate) simply decided boys were easier to market the home PC to (remember, this was decades before the general public took games seriously. People legitmately thought video games were a fad to die out of)

And that stuck. Snowballs into a time where the "nerd" stereotype more or less became the perogative term in the 80's/90's as a quick framing for some undesirable male. Every single piece of media had some stereotype of it, well into the 00's. It's not surprising women were put off. Men were put off too.

No particular mastermind here. Just corporate wanting to make a quick buck off of kids. Those kids just happened to pioneer an entire industry while Hollywood laughed at them.

◧◩
128. dredmo+HG1[view] [source] [discussion] 2023-07-29 09:24:45
>>motoha+uo
What I've noticed about forums is that meta discussions tend to be most prevalent when they're brand new (see for example: <>>363 >), or when there's a major disruption in their operation or management (e.g., Birdsite, presently).

Otherwise ... there's sort of a constant drizzle of mild disappointments and/or outrages, often over moderation, content (posts or commentary) that's objectionable to some faction, and various interfactional skirmishes. HN sees the latter, but even that doesn't seem show much of a long-term trend that I can see. "General news" submissions (by site) have been a major part of front-page submissions from the beginning. Blogs have fallen somewhat, though software projects (identified by GitHub / GitLab URLs) are an increasingly large fraction of posts.

HN has been remarkably even-keeled over the years, without tipping over either into schlerosis, homogeneity, or mass dysfunction, as seems typical of many other online forums I've participated since the late 1980s. I've been looking at various aspects of that through an archive of all front pages from 2007 until a few weeks back (I refresh occasionally, though a few weeks doesn't shift findings much).

◧◩◪◨⬒
130. philsh+gH1[view] [source] [discussion] 2023-07-29 09:30:13
>>dredmo+4E1
Interesting. They had a major refactor of their paywall in August 2019. Maybe that turns people off to posting content that isn’t unique to the NYT. Or maybe paywalled articles are less likely to make the HN front page (?).

https://open.nytimes.com/we-re-launched-the-new-york-times-p...

◧◩◪◨⬒
138. usrusr+NJ1[view] [source] [discussion] 2023-07-29 10:00:56
>>dredmo+4E1
According to the New York Times, no one noticed: https://open.nytimes.com/we-re-launched-the-new-york-times-p...

That cut looks a lot like a consequence of changing approaches to monetarization.

◧◩
141. mschus+kL1[view] [source] [discussion] 2023-07-29 10:20:33
>>knight+4G1
The thing is, technology alone without taking care about ethics and politics as well can be extremely dangerous for society itself. Just look at Facebook, with Zuckerberg's infamous "the dumb fucks trust me" quote from its early days and later on the scandals about Cambridge Analytica or its role in getting Trump elected. Or Whatsapp and the series of lynchings resulting from the new feature of forwarding messages, leading to fake news spreading faster than the old "chain letters".

[1] https://www.theregister.com/2010/05/14/facebook_trust_dumb/

◧◩◪◨⬒
144. JNRowe+wM1[view] [source] [discussion] 2023-07-29 10:32:34
>>johnny+VA1
There was an extension¹ posted a few years ago² which implements blocking on the client side. The possible issue is there may be some users who you wish to block in the general case that do add value in other threads, so you would have to weigh up those decisions.

However, it feels like a bad idea to me, if only because it reduces community building and curation a little. Without a mass of users taking action via flagging or reporting on community problems the end result will surely just tend towards submissions and comment threads that are simply garbage, as that is what new users would see as the social norms.

All that said, I wouldn't blame people for using it. There do seem to be a few members of this community that I'd not want to know in real life, and given their comments in some threads they'd not want to know me either ;)

Edit: I see you mentioned an extension in another thread, and this may be the same one. Still, I'll leave this here in case others want to see the links.

¹ https://github.com/morgante/hn_blocklist

² >>10005699

◧◩◪◨
156. mschus+ZO1[view] [source] [discussion] 2023-07-29 10:54:05
>>Y_Y+WM1
> the clear harms you mention don't seem to me to result from insufficient "diversity"

Of course they do. Simply put: a bunch of young, rich white Western men - they never experience the insane danger that are part of the daily life of women: men stalking them using every way they can, raping them, killing them, or doxxing them just because they can "for the lulz". Both do not experience the racism and discrimination Black people go through, and neither of these have the experience of people in the Balkans, India or other countries with ethnic and religious tensions. And that's just the tip of the iceberg - discrimination runs rampant across all our societies, with technology lessening the effects of some of it (e.g. AI-generated descriptions of images for the blind), and making others exponentially worse (especially online harassment).

More diversity in anything tech automatically means more eyeballs on how a new (or existing) technology can be used by malicious actors to cause harm. For me the worst case in the last few years where this was clearly not done were AirTags - they are immensely useful, but it took over two years and uncountable reports of AirTags being used to facilitate crime of all kind to get a detection feature in Android [1].

[1] https://blog.google/products/android/unknown-tracker-alert-g...

◧◩◪◨⬒⬓⬔
163. dredmo+MW1[view] [source] [discussion] 2023-07-29 12:04:06
>>denton+aN1
That's one hypothesis I'm considering.

It'd be interesting to, say, look at a number of sites which have gone paywall and see how that impact on HN front-page posts.

Off the top of my head, some of those would be:

- NYTimes

- WSJ

- Quora

- WaPo

- LA Times

If anyone has a handy list, especially with dates, I'd appreciate it.

Here's the top 40 "general news" sites with barplots by year. I know that NYT, WSJ, WaPo, LA Times, telegraph.co.uk, and possibly a few others have paywalls and may have implemented them over this period. Pastebin to spare readers here another monster text post, expires in a month:

<https://pastebin.com/raw/nmNXyE8G>

◧◩◪◨⬒⬓⬔⧯
168. DonHop+822[view] [source] [discussion] 2023-07-29 12:45:34
>>dredmo+PR1
Het Nymphéas -bekken - Claude Monet - Valuta van 50 euro of - 1/4 oz Be - 2022

https://www.pieces-et-monnaies.com/nl-nl/products/le-bassin-...

◧◩◪◨⬒⬓⬔⧯
184. dredmo+qy2[view] [source] [discussion] 2023-07-29 16:21:09
>>dredmo+MW1
Paywall dates:

- NY Times: ~August 2019

- BBC: none

- The Guardian: none

- Washington Post: June 2013. <>>5829206 > HN traffic actually rose. Tightened markedly in 2018: <https://web.archive.org/web/20171213135245/https://reason.co...>

- Reuters: April 2021 <>>26820053 >

- NPR: none

- CNN: none

- Slate: 2015 (International readers) <>>9821492 >

- Vice: none?

- LA TImes: Paywalled, 2012. <https://www.huffpost.com/entry/la-times-paywall_n_1299997>

- CNet: none

- Yahoo: none

- SFGate: 2015 <https://old.reddit.com/r/bayarea/comments/2sj78h/death_spira...>

- cbc.ca: none

- CNBC: none

- guardian.co.uk: none

- vox.com: none (though discussed)

- salon.com: none?

Mixed bag on impacts, though I suspect paywalls going up or tightening has a lot to do with FP story trends.

◧◩◪◨⬒⬓⬔⧯
199. lapcat+A83[view] [source] [discussion] 2023-07-29 20:04:23
>>soulof+x53
> It sounds like you didn't want people to engage your comment at all.

Untrue. The replies by bachmeier >>36916075 dredmorbius >>36918463 and philwelch >>36916010 were fine. Respectful, not condescending. The problem is not the existence of replies, it's the content and tone of replies. There's a reason I only mentioned the two of you.

◧◩◪◨⬒⬓
201. dang+2b3[view] [source] [discussion] 2023-07-29 20:19:40
>>DonHop+642
My bottom 10, starting from the bottom:

>>10877423 (Jan 2016 - maybe before we figured out how to treat the dreaded title fever, which drives men mad like mosquitoes in the old northwest)

>>7611005 (April 2014 - what a time warp - remember when anything anti-Elon would get flamed?)

>>8477279 (Oct 2014 - thank god I changed it back or lord knows how bad it couldve gotten)

>>14248635 (May 2017 - everyone wants more monthly threads until the front page fills up with them, guess how you'd like that now)

>>11608112 (May 2016 - fair play for airbnb? how dare i?)

>>10564079 (Nov 2015 - ugh. hn must have gotten better about religious flamewar because that one made me cringe)

>>17780480 (Aug 2018 - i probably wouldn't do that now - but please come back, pvg)

>>8759235 (Dec 2014 - i'm sorry don't hit me!)

>>13752227 (Feb 2017 - oho! we can detach things!)

>>8809021 (Dec 2014 - i have sweeter ways of making the exact same point now)

◧◩◪◨⬒⬓⬔⧯▣▦▧▨
207. dredmo+XM3[view] [source] [discussion] 2023-07-30 00:45:50
>>soulof+Sj3
People do read comments and criticisms about themselves strongly negatively.

This is particularly the case where there seems to be an existing communications failure --- talking cross-purposes, not being heard/seen, failure to acknowledge points or caveats, uncharitable constructions, and the like.

The knack of delivering criticism without aggravation or raising hackles is a true skill. It's one dang's developed well, though he's not always successful. Part of his toolkit is having a set of standard responses to situations which have been honed over time. I'll often link to his own admonitions, though even that doesn't always work, see: <>>36890303 >.

I give an honest effort, then try as best I can to roll with it.

On some platforms it's possible to mute or block profiles, and I find that that's a tremendously useful capability. (I feature the advice as one of the pinned toots on my Fediverse profile: "Block fuckwits". And acknowledge that, yes, sometimes I'm the fuckwit.)

HN for various reasons doesn't offer this as basic functionality. It does have the option of collapsing threads (the setting is retained until a post is about a week or two old, due to systems limitations of the HN server itself). But at least the annoyance isn't staring you in the face whenever you return to your past threads for that week or so.

Even with technologically-mediated options, the practice goes against fundamental human psychology, which is to defend oneself.

(There's a parallel to this I'd thought of in an earlier response but didn't mention at the time: under US law, there is a legal protection against self-incrimination, but that is instituted insidiously in that in order to invoke it, one must forgo the option of defending oneself in the moment. Legally, to have that option one must have a legal advocate, though that costs money, which many defendants in either criminal or civil cases don't have the wealth to access. In that light, Fifth Amendment protections are something of a cruel joke against psychology.)

◧◩◪◨⬒⬓
212. arp242+7r5[view] [source] [discussion] 2023-07-30 16:20:32
>>johnny+IA1
Here's my HN block extension for Firefox: https://gist.github.com/arp242/3159b9cbbbe148c786e1ca10adaae...

Also includes some other things like changed styling and whatnot; it's not really "for publication" and some tweaking may be required, but here it is. Load it manually via about:debugging.

You need to load a post by clicking on the date, and then you can click "bozo" or "block": the "bozo" just marks the post as someone being a "bozo" but doesn't block it. This is useful because everyone can have a bad day or whatever, and that's fine. It's people who consistently seem to be having "bad days" that are the problem – unfortunately there's a small group of highly prolific posters that I find consistently unpleasant, and with just ~20-30 people hidden like this (some of whom really ought to be banned IMHO) I found HN becomes a significantly better experience.

The main problem I have with this is that I can no longer flag or rebuff their posts (whichever may apply), so these people become the proverbial "missing stairs" if everyone starts doing it :-/

[go to top]